Prof. HP Dietz, Sydney Medical School Nepean, University of Sydney, Nepean Hospital, Penrith, NSW 2750, Australia. Email email@example.com
Please cite this paper as: Dickie K, Shek K, Dietz H. The relationship between urethral mobility and parity. BJOG 2010;117:1220–1224.
Objective Urethral mobility is associated with stress urinary incontinence (SUI) and urodynamic stress incontinence, and this is particularly true for mid-urethral mobility. The purpose of this study was to determine whether there is a significant relationship between segmental urethral mobility and vaginal parity in women undergoing urodynamic testing for prolapse or lower urinary tract symptoms.
Design Retrospective study.
Setting Tertiary referral service for multichannel urodynamic testing.
Population Women undergoing urodynamic testing for lower urinary tract symptoms or pelvic organ prolapse.
Methods The stored 3D translabial ultrasound volume data sets of 648 women were assessed. Measurements were performed using post-processing software in volumes obtained at rest and on maximal Valsalva manoeuvre. Analysis was based on a co-ordinate system using the dorsocaudal margin of the pubic symphysis. The urethral length was traced and divided into five equal segments. Mobility vectors are determined by the formula √[(yV − yR)2 + (xV − xR)2], where V indicates Valsalva and R indicates rest, with ‘x’ as the vertical distance and ‘y’ as the horizontal distance from the dorsocaudal margin of the pubic symphysis.
Main outcome measures Mobility vector lengths.
Results The distal urethra is consistently the least mobile part of the organ, regardless of parity. Vaginal childbirth seems to increase urethral mobility by about 20% for all urethral segments (all P ≤ 0.009). The first vaginal delivery showed the greatest effect, particularly on mid-urethral mobility.
Conclusions There is a significant association between urethral mobility and vaginal delivery in women seen for symptoms of pelvic floor dysfunction, affecting all segments of the urethra equally. Most of this effect seems to result from the first vaginal birth.
In stress urinary incontinence (SUI), urine leaks from the urethra when bladder pressure exceeds intra-urethral pressure in the absence of detrusor activity. When abdominal pressure rises, such as on coughing, sneezing, laughing, running, jumping and with sexual activity, the vesical pressure increases, resulting in involuntary leakage. SUI accounts for at least half of all cases of urinary incontinence.1 It is often (but not always) associated with bladder neck hypermobility, and pregnancy and childbirth have been shown to increase the mobility of the bladder neck.2–7 This may be the result of congenital factors, hormonal changes or actual trauma to urethral support structures. A definitive mechanism of pathogenesis remains to be elucidated.
Recently, the focus of research into the pathophysiology of stress urinary incontinence has shifted from the bladder neck to the mid-urethra, partly because of the superior curative effect of mid-urethral slings. We have recently developed a new methodology to quantify segmental urethral mobility.8,9 Urethral mobility is an interesting topic for study as damage to pubo-urethral ligaments and the levator ani during vaginal childbirth may alter the support of the urethra and urethrovesical junction, thereby increasing the mobility of the organ. This may impair pressure transmission to the urethra, and therefore closure during periods of increased intra-abdominal pressure.10
The senior author’s unit has developed a method to determine segmental urethral mobility, the ‘urethral motion profile’ or UMP. This method has previously been used in women before and after childbirth.8 In that study it was concluded that there is a significant relationship between urethral mobility and childbirth, particularly after instrumental delivery, with changes in urethral support affecting the entire urethra.8 In the current project we aimed to investigate the relationship between urethral mobility and parity in a large cohort of women complaining of symptoms of pelvic floor and lower urinary tract dysfunction.
In this retrospective study, a total of 648 women were seen at a tertiary referral service for multichannel urodynamic testing (Neomedix Acquidata, Sydney, Australia) and 4D pelvic floor ultrasound imaging (GE Kretz Voluson 730 Expert, GE Medical, Ryde, NSW, Australia) between December 2006 and September 2009. Of those, 163 were excluded because of previous prolapse or anti-incontinence surgery. In 29 cases, ultrasound volume data was unavailable because of clerical error or data corruption, or was technically inadequate for UMP determination. All subsequent data refers to the remaining 455 complete data sets. Volume imaging data was obtained in the supine position while at rest, and on maximal Valsalva manoeuvre following bladder emptying. At least three Valsalva manoeuvres were attempted, with the most effective Valsalva being used for the analysis.11 The Voluson-type systems utilise the rapid oscillation of an array of piezoelectric elements to obtain volume data, producing a fan-shaped block of data several times per second. This allows the assessment of changing anatomy (e.g. during a Valsalva) at a temporal resolution of several hertz.11
Four-dimensional imaging also has the advantage that representations of entire manoeuvres (e.g. 36 volumes obtained over the course of 5–20 seconds) can be saved as volume data and stored on a server, awaiting analysis. This simplifies retrospective analysis as measurements can be obtained at a later date, without inconvenience to the patient, and with much greater ease for the operator. Of course measurements such as those for a urethral motion profile could be obtained with much simpler 2D systems, and live, but it would take much more time and require the patient to stay while the data is obtained. This advantage applies not just for relatively simple analysis, such as that of urethral mobility in the midsagittal plane, but also to more complex tasks such as the evaluation of reflex muscle activity, of organ descent over time, or volume measurements of complex 3D shapes.
The UMP analysis was performed using 4D View v5.0 (GE Medical, Ryde, NSW, Australia) on a desktop PC, and was blinded against all clinical data. 4D View is similar to the dicom viewer software used in radiological imaging, with the added ability to manipulate cine loops of 3D volume data. Distance, area and volume measurements can be performed in any arbitrarily defined plane, and in semi-transparent volumes.
Urethral length was traced and divided into five equal segments by six equidistant points along the trace length. Point 1 is at the internal urethral meatus and point 6 is at the external meatus, using the dorsocaudal margin of the pubic symphysis as the point of reference (see Figure 1). Utilising an Excel macro, a semi-automated measurement algorithm produces x and y coordinates on a bitmap imported from 4D View. Mobility vectors for each point are determined by the formula √[(yV − yR)2 + (xV − xR)2] with ‘x’ as the vertical distance from the dorsocaudal margin of the pubic symphysis and ‘y’ as the horizontal distance from the dorsocaudal margin of the pubic symphysis at rest (R) and on Valsalva (V).8
Statistical analysis was performed using Minitab v.13 (Minitab Inc., State College, PA, USA) after testing for normality (Kolmogorov–Smirnov testing). We used anova analysis and two-sample Student’s t test statistics to test segmental urethral mobility against parity.
Repeatability measures of urethral coordinates were excellent, as determined in a test–retest series of 20 UMPs (120 mobility vectors), with an intraclass correlation of 0.78 (95% CI 0.63–0.87) for segmental urethral mobility. The mean age of the 455 patients was 55 (range 20–90). They presented with stress incontinence (n = 362, 80%), urge incontinence (n = 325, 71%), frequency (n = 134, 29%), nocturia (n = 198, 44%), symptoms of voiding dysfunction, including hesitancy, poor stream, stop and start voiding (n = 105, 23%), and symptoms of prolapse, including vaginal lump or a dragging sensation (n = 184, 40%). In total, 415 women had delivered vaginally (91%) and 124 (27%) had had a previous hysterectomy. On clinical assessment 58% (n = 264) had a significant prolapse [International Continence Society pelvic organ prolapse quantification system (ICS POP-Q) stage 2 or higher). There were 195 (21%) cystoceles, 35 (8%) uterine prolapses, 18 (4%) enteroceles and 166 (36%) rectoceles. Levator avulsion was diagnosed in 18% of the women (n = 84).
In two cases urodynamic testing was not completed, one because of urethral stenosis and one because of patient refusal, leaving 453 data sets. Urodynamic stress incontinence (USI) was diagnosed in 309 cases (68%), detrusor overactivity (DO) was diagnosed in 116 cases (26%), and voiding dysfunction (VD) was diagnosed in 129 cases (28%). The latter was defined as an abnormal pressure flow study and/or residual of over 50 ml on free flowmetry, and/or a maximum flow rate under the tenth centile of the Liverpool nomogram.12 Free flowmetry was obtained for 376 patients (i.e. n = 453 − 77). The mean voided volume was 248 ml, and the mean maximum flow-rate centile was 33 (range 0–98). The median residual was 0 (interquartile range 0–20).
All ultrasound data in this analysis was normally distributed. The vectors were measured at a mean of 2.96, 2.53, 2.13, 1.84, 1.76 and 1.81 cm for points 1–6, respectively, indicating that the distal urethra is generally less mobile than the proximal. There was significantly higher mobility throughout the entire urethra in vaginally parous women (Table 1). The first vaginal delivery was shown to have the greatest effect on all vector lengths, as seen in Figure 2. On multivariable analysis, vaginal operative delivery was found to be an effect modifier in this relationship (adjusted R2 = 0.06, P = 0.02), in the sense that it increased mobility.
Table 1. A comparison of segmental urethral mobility between vaginally nulliparous and vaginally parous women (two-sample Student’s t-test). Figures signify means in cm, with standard deviations in brackets
Vaginally nulliparous (n = 40)
Vaginally parous (n = 415)
Age at first delivery (mean 23 years, range 12–40 years) was also analysed, but was not shown to be an effect modifier. Age at presentation was associated with urethral mobility in a complex form, with an increase in mobility until the age of menopause, and a reduction later in life, and this was the case for both the bladder neck (point 1) and the mid-urethra (point 4) (see Figure 3).
It is generally accepted that vaginal childbirth is associated with an increased risk of SUI in later life.1,13 Urethral mobility is considered to be a significant factor in the pathogenesis of SUI, but until recently the available imaging technology only allowed the assessment of bladder neck mobility, rather than the mobility of the entire urethra. Whereas magnetic resonance imaging theoretically has a higher resolution, it is clearly inferior to ultrasound in assessing dynamic changes. Translabial or transperineal ultrasound is the imaging method of choice in assessing urethral mobility, and recent technical advances have improved tissue discrimination so much that the external meatus is routinely identifiable. The Four-Dimensional imaging allows the registration and storage of volume ultrasound data at high spatial and temporal resolutions,11 conveniently enabling later analysis by an operator blinded against all clinical data. Using 4D ultrasound volume data, we have developed a methodology that allows the assessment of segmental urethral mobility,8 allowing us to explore its association with parity in a large group of symptomatic women seen for urodynamic testing.
The anatomical structures influencing urethral mobility have been a point of contention for many years. Krantz first described the ‘pubourethral ligaments’, which are supposed to anchor the proximal and central urethra to the pubic rami.14 Magnetic resonance imaging data15 and our own cadaver dissections confirm the existence of such structures, even if their structure and location are highly variable between individuals. However, there are other theories of urethral support. De Lancey’s ‘hammock hypothesis’ states that a connection exists between the vaginal wall and endopelvic fascia to the arcus tendineus fasciae pelvis and levator ani muscles,16 and that this hammock is essential for the preservation of continence. Remarkably, other authors have claimed that there are no identifiable ligamentous structures fixing the urethra to the pelvic sidewall.17
Urethral fixation (by whichever mechanism) is likely to be important for urinary continence because any ‘fixation’ or ‘tethering’ implies force (or pressure) transmission. In a subset of the population tested here we have previously been able to show that urethral mobility, as determined by UMP, is strongly associated with urodynamic stress incontinence, but not with other urodynamic diagnoses.9
Regardless of the nature of structures that tether the urethra to the os pubis and/or pelvic sidewall, it is clear that such structures must exist. The part of the vaginal wall that overlies the distal urethra is the least likely to prolapse, even in women with complete vault eversion, a fact that has been known for more than a hundred years.18 In general, the distal urethra is the least mobile part of the organ, implying fixation, and this clinical observation has been confirmed using the methodology employed for this study.8 In this large series we again found that the distal urethra is the least mobile part of the organ, regardless of parity.
Our hypothesis was that ‘vaginal parity is positively associated with segmental urethral mobility, as measured by the urethral motion profile’. This hypothesis was found to be true. The difference between vaginally nulliparous and vaginally parous women was consistently in the order of about 20%, affecting all parts of the urethra similarly. Most of the effect seems to result from the first vaginal delivery, especially for the mid and distal urethra. This closely agrees with own data comparing urethral mobility in late pregnancy and after childbirth.8 Vaginal operative delivery seems to have only a minor effect on this relationship. This is in accordance with previous epidemiological studies.19
Limitations of this study include the fact that the women of this data set are unlikely to be a true representation of the general population as they all attended urodynamic testing for symptoms of bladder dysfunction and/or prolapse. It may not be possible to generalise our findings to the general population. A second limitation relates to the reproducibility of our data in measuring urethral mobility on translabial ultrasound. Test–retest data was obtained from stored ultrasound cine volume data sets. It must be recognised that recalling the patient for a second assessment may result in a higher test–retest variability. Despite these limitations, our study constitutes strong evidence that vaginal childbirth affects urethral mobility, and that this change is largely caused by the first vaginal delivery.
There is a significant association between urethral mobility and vaginal delivery in women seen for symptoms of pelvic floor dysfunction, and this seems to affect all segments of the urethra equally. Most of this effect seems to result from the first vaginal birth.
Disclosure of interests
HPD has acted as a consultant for American Medical Systems (Minnetonka, MN, USA) and Continence Control Systems (Sydney, Australia), has accepted speaker’s fees from GE Medical Ultrasound (Sydney, Australia), American Medical Systems, and Astellas (Tokyo, Japan), and has benefited from equipment loans provided by General Electric, Bruel and Kjaer (Gentofte, Denmark), and Toshiba (North Ryde, Australia).
Contribution to authorship
KJD was responsible for data acquisition and writing the manuscript. KLS trained KJD, supervised the data acquisition, and proofread the manuscript. HPD planned the study, supervised the data acquisition, performed statistical analysis, and proofread the manuscript.
Details of ethics approval
Sydney West Area Health Service Human Research Ethics Committee, reference number: 09/42.
We acknowledge the help of Adrienne Kirby, a biostatistician, with data analysis, and that of Athina Pirpiris who participated in the UMP analysis.